Doctors Are Using More AI. What Makes Them Trust It?

Artificial intelligence has become an integral part of physicians' daily work. A recent AMA survey found that 81% of physicians now use AI in their practice, up from previous years. While more than 75% believe AI improves patient care, concerns about safety and data privacy persist. Eighty-eight percent of physicians deem robust validation essential, along with 86% prioritizing data privacy.

Clear liability frameworks also rank high among the regulatory steps needed to boost trust in AI technology. As AI becomes more prevalent, clinicians must understand when and why they can trust its recommendations, with trust and transparency closely linked in the eyes of healthcare leaders. Dr. Matt Sakumoto, chief clinical product officer at Nabla, emphasized that transparency drives trust as long as clinicians have evidence about the tools' performance and data usage.

Theo Zanos of Northwell Health argued that sustained performance matters more than transparency for building physicians' trust. However, research on explainable AI shows mixed results, as simply providing explanations does not automatically increase trust. Patients prioritize AI performance, as a recent JAMA Network Open study found it to be the most significant factor in determining preferences for AI-assisted medical encounters.

Health systems must make AI evaluation an ongoing process, prepared to stop, reduce usage, or pivot if reliability, privacy, or safety concerns arise. Transparency for health insurer Elevance Health translates to explainability, as they centralize AI development and subject applications to a responsible-AI framework. Ultimately, clinicians should remain involved in AI adoption decisions and continuously evaluate tools' reliability as data and models evolve, with accountability shared between clinicians and organizations.